Maintenance represents a large part of the total operational costs of vehicle fleets. We perform reliability analysis based on a dataset of maintenance records from the fleet of light commercial vehicles used by BT Fleet Solutions and their customers (Post Office, National Grid, etc.). In our preliminary analysis, we first verify that reliability deteriorates with age, mileage, and the number of historical maintenance activities, as we also find heterogeneous behavior associated with different geographic locations and vehicle makes. Drawing on this background, we propose a dynamic maintenance policy. We build it on a data-driven method, Conditional Inference Trees (CIT), which provides a good balance between applicability (with loose assumptions) and interpretability (being a “white-box” solution). Since there is no closed-form solution for maintenance costs based on tree-structured reliability, we develop a simulation-based method to estimate the maintenance cost for the CIT-based maintenance policy, along with a new backward optimization procedure for the operational parameters. Based on the cost analysis, we demonstrate the effectiveness of the CIT-based maintenance policy in comparison to two prevalent maintenance policies. Finally, we identify opportunities for further improvements in the operations of the company, and for maintenance operations in general.

We study an inventory routing problem with time windows (IRPTW). A single supplier serves a set of customers to fulfill their demand throughout a planning horizon. Each customer can be visited only within designated time windows, assuming each customer provides a single delivery time window valid for all periods. We propose a two-commodity flow formulation for this problem and develop a branch-price-and-cut algorithm to solve it. We test our algorithm on a rich set of benchmark instances with fixed and time-varying demands and with up to 100 customers. Computational experiments demonstrate the effectiveness of this approach. Within a 2-hour time limit, it can provide a lower bound for all 492 tested instances. Furthermore, 211 of them are solved to optimality, achieving an average optimality gap of 4.07% for large instances and closing the gap for 11 instances for the first time in the literature.

The two-echelon inventory-routing problem (2E-IRP) addresses the coordination of inventory management and freight transportation throughout a two-echelon supply network. The latter consists of geographically widespread customers whose demand over a discrete planning horizon can be met from their local inventory, or from intermediate facilities’ inventory. Intermediate facilities are located in the city outskirts and are supplied by distant suppliers. Assuming a vendor-managed inventory system, the 2E-IRP aims to minimize transportation and inventory costs while meeting customers’ demands. To solve this problem, we propose a route-based formulation and develop a branch-and-price algorithm. A labeling algorithm solves one pricing subproblem for each combination of time period and intermediate facility. We generate 400 instances and obtain optimal solutions for 149 of them. We provide an upper bound for another 77 instances with a gap of less than 5% (with an average of 2.79%) and an upper bound for the rest of the instances with an average gap of 11.33%. We provide comprehensive analyses to evaluate the performance of our solution approach.

Effective coordination of operational decisions in today’s global supply chains has grown increasingly important. This paper considers a two-echelon inventory-routing problem under a vendor-managed inventory system, where suppliers are responsible for fulfilling the demands of geographically scattered customers through a set of intermediate facilities over a finite planning horizon. The problem involves determining the routing and delivery decisions to minimize the total routing and inventory costs. We introduce an effective two-phase matheuristic approach that combines tabu search and mathematical programming models. Computational experiments show that our approach achieves excellent results regarding solution quality and computational time. For small instances, the matheuristic finds 99 optimal solutions out of 165 known optimal solutions and achieves an average gap of (−2.32%) over 235 instances with a known best upper bound only, improving 159 known best upper bounds. Our approach also solves larger instances within a reasonable computational time and provides upper bounds for all 400 large-sized instances for the first time in the literature. Through a comprehensive series of experiments, we provide insights into the efficacy of different components of the proposed solution method.

Closed-loop supply chains (CLSCs) are seen as one of the circular economy’s leading approaches for reducing our natural environment load. Many CLSC models require collaboration among different parties. Game theory (GT) offers a way to consider the profits of all parties in a CLSC, providing insight into the costs and benefits to the involved parties in an objective and quantitative way. Presently, available reviews on the use of GT, in the context of CLSC, are quite limited and consider only a few relevant elements. Here, we present a new and more extensive framework, focusing on the collaboration structure of CLSCs. It contains a content-based analysis of 230 papers based on a four-step systematic literature review process. The characteristics studied are channels for collection, reprocessing and selling, the planning horizon, and the types of games. The structures found are graphically reviewed, leading to 196 different structures. The results show that, so far, most attention has been paid to the dual-channel collection, where collection by two retailers (dual-retailer) is the most studied case. With respect to selling, most attention has been paid to situations with two selling channels (dual-selling), i.e., one channel managed by a manufacturer and one channel managed by a remanufacturer. Studies have prioritized the role of manufacturers as that of the leader and collector. Finally, a number of directions for further research are pointed out.

Sustainable supplier selection is a pressing mater in supply chain management. This paper tackles the main pain points in decision making during supplier selection and order allocation (SSS & OA) under uncertainties for a multi-item, multi-period setting, where each supplier has its own pricing policy. In the first phase, a hybrid BWMER method is employed for evaluating and ranking suppliers. In this method, the best worst method (BWM) for determining weights of the sustainability criteria and evidential reasoning (ER) for evaluating suppliers under uncertainty are used. Based on constraints related to demand, capacity, inventory, and allowed shortages in the next phase, a bi-objective mathematical model is presented, to make a trade-off between sustainability and economic cost. The demand is assumed to be stochastic, and there are uncertainties in the availability of suppliers during different periods. The combination of these two uncertainties is studied via a set of scenarios. A new integrated solution approach based on stochastic programming and dynamic programming to solve the biobjective model under uncertainties is presented. The proposed approach results are compared with the results obtained via revised multi-choice goal programming and the Epsilon constraint method for a real-life case. These comparisons show that by using the new method, better and faster results are obtained. Sensitivity analysis is used to reveal the effect of quantity discounts, uncertainties in suppliers' availability and demand. Results obtained for a real-life case study are presented and discussed.

A heuristic substitutions policy to control inventories for a hybrid manufacturing/remanufacturing system with stochastic demand considering downward substitutions between three markets (new/once remanufactured/twice remanufactured units of one product) is presented. The policy governs whether or not to allow substitution options between the markets, as well as the related quantities, where the behavior of customers related to the substitution options is explicitly taken into account. The main aim of substitution, apart from immediately fulfilling demand, is to manage the availability of returns to improve the net profit and the fill rates. The values of the policy parameters are determined via simulation-based optimization. To justify the use of the heuristic policy, a simplified Markov decision process to obtain the optimal policy is provided. The performance of the heuristic policy is examined via numerical experiments. The results show that profitability and serviceability can be improved using the proposed policy, even if the actual substitutions result in a direct loss. It is also shown that substitution is not useful if the used product holding costs are beyond a certain value. Exceeding this value may even give rise to no longer remanufacture. Investigation of the effect of return rates reveals that higher return rates result in a higher net profit and higher serviceability. Further, the importance of making customers willing to accept substitution options is shown. Suggestions for further research are indicated.

In response to the rapid growth of environmental problems related to air transportation including emissions and noise, aviation authorities and industries have implemented stricter environmental regulations and targets in the early 2000s to encourage airlines to become greener. Different emission mitigation measures to achieve these targets have been developed. This study introduces a new mitigation measure. In addition to minimizing the total cost of flow and establishing hubs, the measure aims at minimizing the greenhouse gas emissions, fuel consumption and noise in the design of airline hub and-spoke network. We develop a multi-objective mixed integer-programming model and use several methodologies to determine the best design. Our computational results are based on the CAB data set for the domestic US aviation sector. For assessing the value of the new mitigation measure, we develop cumulative marginal abatement cost curves for reducing the projected annual CO2 emissions from 2020 to 2050 for the domestic US air transportation. Our results indicate that using the new measure can cost-effectively decrease the projected cumulative (2020-2050) CO2 emissions by more than 250 and 200 million tonnes relative to the classical model, which relies on minimizing the total cost of flow and opening hubs only. We conclude that the new measure not only provides a practical solution for the airlines facing high fuel cost, emission trading schemes or carbon tax systems but it can also play an important role in achieving sustainable and environmental-friendly targets. (C) 2019 Elsevier Ltd. All rights reserved.

Condition-based maintenance (CBM) has received increasing attention in the literature over the past years. The application of CBM in practice, however, is lagging behind. This is, at least in part, explained by the complexity of real-life systems as opposed to the stylized ones studied most often. To overcome this issue, research is focusing more and more on complex systems, with multiple components subject to various dependencies. Existing classifications of these dependencies in the literature are no longer sufficient. Therefore, we provide an extended classification scheme. Besides the types of dependencies identified in the past (economic, structural, and stochastic), we add resource dependence, where multiple components are connected through, e.g., shared spares, tools, or maintenance workers. Furthermore, we extend the existing notion of structural dependence by distinguishing between structural dependence from a technical point of view and structural dependence from a performance point of view (e.g., through a series or parallel setting). We review the advances made with respect to CBM. Our main focus is on the implications of dependencies on the structure of the optimal CBM policy. We link our review to practice by providing real-life examples, thereby stressing current gaps in the literature. (C) 2017 Elsevier B.V. All rights reserved.

We propose a new condition-based maintenance policy for complex systems, based on the status (working, defective) of all components within a system, as well as the reliability block diagram of the system. By means of the survival signature, a generalization of the system signature allowing for multiple component types, we obtain a predictive distribution for the system survival time, also known as residual life distribution, based on which of the system's components currently function or not, and the current age of the functioning components. The time to failure of the components of the system is modeled by a Weibull distribution with a fixed shape parameter. The scale parameter is iteratively updated in a Bayesian fashion using the current (censored and non censored) component lifetimes. Each component type has a separate Weibull model that may also include test data. The cost-optimal moment of replacement for the system is obtained by minimizing the expected cost rate per unit of time. The unit cost rate is recalculated when components fail or at the end of every (very short) fixed inter-evaluation interval, leading to a dynamic maintenance policy, since the ageing of components and possible failures will change the cost-optimal moment of replacement in the course of time. Via numerical experiments, some insight into the performance of the policy is given. (C) 2017 Elsevier Ltd. All rights reserved.

For still more companies, it is or will become important to pay attention to the possibilities for reusing the products they produce and the items, like pallets and package materials, that they use for distributing their products or which are used by others for supplying their products to them. One important reason for the above is the growing concern for the natural environment, among others resulting in environmental laws which not only force companies to take back their products from their customers and the items used for the distribution of these products when these products or distribution items (DIs) are no longer desired by these customers, but also to take care of the environmentally friendly disposal of these products and Dis. However, due to the same reason, this disposal is becoming still more difficult and expensive (see e.g., Cairncross, 1990). Apart from being forced by law, companies feel forced to do the above because of competition and public opinion. But there are more reasons why it may be worthwhile for companies to consider reuse: there are products, components, materials and DIs that can be obtained cheaper or more quickly via reuse than via purchasing or producing anew.

A problem that then can arise is called prior-data conflict, see, e.g., [3]: from the viewpoint of the expert, the observed data seem surprising, i.e., the information derived from observed data is in conflict with prior assumptions. Models based on conjugate priors can be insensitive to prior-data conflict, in the sense that the spread of the posterior distribution does not increase in case of such a conflict [see 4, §A.1.2 for two examples], thus conveying a false sense of certainty by communicating that we know the reliability of a system quite precisely when in fact we do not.

We consider the problem of planning preventive maintenance and overhaul for modules that are used in a fleet of assets such as trains or airplanes. Each type of module, or rotable, has its own maintenance program in which a maximum amount of time/usage between overhauls of a module is stipulated. Overhauls are performed in an overhaul workshop with limited capacity. The problem we study is to determine aggregate workforce levels, turn-around stock levels of modules, and overhaul and replacement quantities per period so as to minimize the sum of labor costs, material costs of overhaul, and turn-around stock investments over the entire life-cycle of the maintained asset. We prove that this planning problem is strongly \(\mathcal{NP}\)-hard, but we also provide computational evidence that the mixed integer programming formulation can be solved within reasonable time for real-life instances. Furthermore, we show that the linear programming relaxation can be used to aid decision making. We apply the model in a case study and provide computational results for randomly generated instances.

This paper deals with the coordination of manufacturing, remanufacturing and returns acceptance control in a hybrid production-inventory system. We use a queuing control framework, where manufacturing and remanufacturing are modelled by single servers with exponentially distributed processing times. Customer demand and returned products arrive in the system according to independent Poisson processes. A returned product can be either accepted or rejected. When accepted, a return is placed in a remanufacturable product inventory. Customer demand can be satisfied as well by new and remanufactured products. The following costs are included: stock keeping, backorder, manufacturing, remanufacturing, acceptance and rejection costs. We show that the optimal policy is characterized by two state-dependent base-stock thresholds for manufacturing and remanufacturing and one state-dependent return acceptance threshold. We also derive monotonicity results for these thresholds. Based on these theoretical results, we introduce several relevant heuristic control rules for manufacturing, remanufacturing and returns acceptance. In an extensive numerical study we compare these policies with the optimal policy and provide several insights.

We consider a single-product make-to-stock manufacturing-remanufacturing system. Returned products require remanufacturing before they can be sold. The manufacturing and remanufacturing operations are executed by the same single server, where switching from one activity to another does not involve time or cost and can be done at an arbitrary moment in time. Customer demand can be fulfilled by either newly manufactured or remanufactured products. The times for manufacturing and remanufacturing a product are exponentially distributed. Demand and used products arrive via mutually independent Poisson processes. Disposal of products is not allowed and all used products that are returned have to be accepted. Using Markov decision processes, we investigate the optimal manufacture-remanufacture policy that minimizes holding, backorder, manufacturing and remanufacturing costs per unit of time over an infinite horizon. For a subset of system parameter values we are able to completely characterize the optimal continuous-review dynamic preemptive policy. We provide an efficient algorithm based on quasi-birth-death processes to compute the optimal policy parameter values. For other sets of system parameter values, we present some structural properties and insights related to the optimal policy and the performance of some simple threshold policies. (C) 2013 Elsevier B.V. All rights reserved.

This paper describes a production scheduling model that can be run for an emulsion polymerization (EP) production process. This scheduling model is used to maximize the total profit of the production process by calculating a production plan which optimizes costs and profits. These costs include production, reactor cleaning, component storage and purchasing of raw materials. It is further noted that the model is structured as a State-Task Network. To ensure feasibility we have introduced new constraints for the task assignment. All coding was done in AIMMS.The model is tested with 2 case studies that represent an EP production process (A smaller and a larger problem), in which we have varied the process topology, the number of raw materials and products. In addition a solver comparison was done from which we found that GUROBI can handle the problem most effectively.

In earlier work we have developed and tested a scheduling [4] and planning model [5] in the AIMMS software. In this follow-up contribution we will develop a model that can be used to blend crudes from the storage tanks to the right API's required for the CDU's. This so called control level model optimizes the refinery activities at a shorter time scale than the scheduling model, typically at the minutes to hour scales. We will make a first step to identify the information flow between the scheduling model and the control level model. Especially the development of the crude oil layering constraints in the tanks is brought under the attention. With case studies we further demonstrate how the model works.

This paper deals with the simultaneous acquisition of capacity and material in a situation with uncertain demand, with non-zero lead-times for the supply of both material and capacity. Although there is a lot of literature on the time-phased acquisition of capacity and material, most of this literature focuses on one of the two decisions. By using a dynamic programming formulation, we describe the optimal balance between using safety stocks and contingent workforce for various lead-time situations. We compare the cost ingredients of the optimal strategy with the standard inventory approach that neglects capacity restrictions in the decision. The experimental study shows that co-ordination of both decisions in the optimal strategy leads to cost reductions of around 10%.We also derive characteristics of the optimal strategy that we expect to provide a thorough basis for operational decision making.

Cores acquired by a remanufacturer are typically highly variable in quality. Even if the expected fractions of the various quality levels are known, then the exact fractions when acquiring cores are still uncertain. Our model incorporates this uncertainty in determining optimal acquisition decisions by considering multiple quality classes and a multinomial quality distribution for an acquired lot. We derive optimal acquisition and remanufacturing policies for both deterministic and uncertain demand. For deterministic demand, we derive a simple closed-form expression for the total expected cost. In a numerical experiment, we highlight the effect of uncertainty in quality fractions on the optimal number of acquired cores and show that the cost error of ignoring uncertainty can be significant. For uncertain demand, we derive optimal newsboy-type solutions for the optimal remanufacture-up-to levels and an approximate expression for the total expected cost given the number of acquired cores. In a further numerical experiment, we explore the effects of demand uncertainty on the optimal acquisition and remanufacturing decisions, and on the total expected cost.